Open Internet by MindsNet

Addressing Age Prediction Imbalance with Limited Datasets

The author faces a challenge in predicting age from voice/sound data due to an imbalanced dataset with limited samples for older ages and no data below 10 and above 90. The existing dataset is labeled in age ranges (e.g., teen, twenties) rather than precise ages, making it difficult to treat age as a regression problem. The author seeks a solution to adapt to these limitations.

Computing & Technology, Computer Science, Machine Learning